ASO Skills
AI agent skills for App Store Optimization, keyword research, metadata, competitors and app growth.
Website
Replay-based evals for AI agents — traces you can run, not just read.
Built with love by Shushant Lakhyani. Building bestproductivityresources.com
Screenshot or bookmark this page so you find it later, then share it with your network.
Kitaru turns production agent traces into reproducible regression tests. Wrap your agent once, or import existing traces from Langfuse, LangSmith, Braintrust, Logfire or Arize Phoenix, then replay any session against your next model, prompt or code change. Replays re-execute your real code with recorded tool-call responses, so an unchanged run gives a faithful baseline and a forked run shows exactly what changed. Human reviewers can pin judgments to exact trace locations, calibrate evaluators, freeze cohorts and run experiments before shipping. It is open source, self-hosted, ships with a Python SDK and TypeScript client, an MCP server, and coding-agent skills for Claude Code and Cursor.
AI engineering teams and agent builders who want regression testing and evals tied to real production traces.
Teams without existing agent traces or those that only need simple prompt-unit tests.
Similar resources to Kitaru, picked from the directory.
AI agent skills for App Store Optimization, keyword research, metadata, competitors and app growth.
Claude Code skills that turn a startup idea into a validated strategy in one afternoon.
The model-definition framework for state-of-the-art machine learning.
Give your coding agent a video editor — local FFmpeg, no cloud, no API keys.
Open-source antivirus for AI agents: block risky tools, secret access, prompt injection and malicious packages.
Token-efficient AI agent — same budget, higher intelligence density.
Free resources that cover similar ground.
MCP server that orchestrates language servers into agent-native workflows.
Local-first MCP gateway — one port for every tool and AI client, with lazy discovery saving ~90% of tokens.
A high-performance multithreaded task execution engine for agents and dweb apps.
A constitutional governance layer for consequential AI actions.
Open-source AI security verification for models, endpoints, MCP servers and agent traces.
The open-source MCP server that lets AI agents 3D print.
Made Kitaru? Add this free badge to your website or README. It links to this page.
<a href="https://bestproductivityresources.com/resources/kitaru" target="_blank" rel="noopener"><img src="https://bestproductivityresources.com/badges/featured-light.svg" alt="Featured on Best Productivity Resources" width="250" height="54" /></a>Open stats
This site's analytics are open for everyone to see.